Phosphorylation of serine, threonine and tyrosine residues by protein kinases represents one of the most common post-translational regulatory modifications of proteins. More than 200 protein kinases have been described, following either purification to homogeneity or molecular cloning (see, Hunter, T. (1991), Methods Enzymol., 200:3-37; Hanks, S. K., et al. (1991), Methods Enzymol., 200:38-81; Hanks, S. K. 1991), Curr. Opin. Struct. Biol., 1:369-383; and Hubbard, M. J., et al. (1993) Trends Biochem. Sci., 18:172-177). It is thought that as much as 2-3% of eukaryotic genes encode protein kinases. The importance of protein kinases in physiological processes has stimulated an active search for specific inhibitors with potential pharmacological interest (see, Hidaka, H., et al. (1992), Annu. Rev. Pharmacol. Toxicol., 32:377-397). Several classes of compounds have been identified, such as staurosporine, naphthalene sulfonamides (W7, ML-9, SC-9), isoquinoline derivatives (H-7, H-8, KN-62), sphingosine, tyrphostins and others, but in most cases these inhibitors display broad specificity. Only some pseudosubstrate autoinhibitory peptides show a high degree of specificity.
Cyclin-dependent kinases (CDK), in particular, have recently raised considerable interest in view of their essential role in the regulation of the cell division cycle (CDC) (see, Nigg, E. A. (1993), Trends in Cell Biol., 3:296-301; and Sherr, C. S. (1993), Cell, 73:1059-1065). CDKs are highly conserved among eukaryotic species. Higher eukaryotic cells contain several isoforms of CDKs that become activated in specific phases of the cell cycle. CDKs consist of a catalytic subunit, the prototype of which is CDC2, and a regulatory subunit (cyclin). Six human CDK proteins have been described so far (see, Meyerson, M., et al. (1992), EMBO J., 11:2909-2917; Meyerson, M., et al. (1994), Mol. Cell. Biol., 14:2077-2086; and Van den Heuvel, S., et al. (1993), Science, 262:2050-2054), namely, CDK1 (also known as CDC2) and CDK2-6. With the exception of CDK3, for which the regulatory cyclin has not yet been identified, all these CDKs proteins are regulated by the transient association with one member of the cyclin family, i.e., cyclin A (CDC2, CDK2), B1-B3 (CDC2), D1-D3 (CDK2, CDK4, CDK5, CDK6), E (CDK2). Each step of the cell cycle is thought to be regulated by such CDK complexes: G1/S transition (CDK2/cyclin E, CDK3/unknown cyclin, CDK4/cyclin D1-D3, CDK6/cyclin D3), S phase (CDK2/cyclin A), G2 (CDC2/cyclin A), G2/M transition (CDC2/cyclins B).
CDKs are able to phosphorylate many proteins that are involved in cell cycle events, including histones, lamins and tumor suppressor proteins, such as the retinoblastoma gene product pRb (see, Norbury, C., et al., supra, Matsushime, H., et al. (1992), Cell, 71:323-334, Nigg, E. E. (1993), Curr. Opin. Cell. Biol., 5:187-193). In accordance with their central role in the cell cycle, enzyme activity is tightly controlled by multiple mechanisms. Kinase activation requires complex formation with regulatory cyclin proteins as described above, followed by an activating phosphorylation on Thr-161 in CDC2 or the corresponding Thr in the other CDKs (see, e.g., Gould, K. L., et al. (1991), EMBO J., 10:3297-3309; Desai, D., et al. (1992), Mol. Biol. Cell, 3:571-582; Solomon, M. J., et al. (1992), Mol. Biol. Cell, 3:13-27). In addition, enzyme activity is negatively regulated by phosphorylations at Tyr-15 and/or Thr-14 (see, e.g., Solomon, M. J., et al., supra; Gu, Y., et al. (1992), EMBO J., 11:3995-4005; Krek, W., et al. (1991), EMBO J., 10:3331-3341; Norbury, C., et al. (1991), EMBO J., 10:3321-3329; Parker, L. L., et al. (1992), Proc. Natl. Acad. Sci. U.S.A., 89:2917-2921; McGowan, C. H., et al. (1993), EMBO J., 12:75-85), or by complex formation with inhibitor proteins like p16 (see, Serrano, M., et al. (1993), Nature (London), 366:704-707; Kamb, A., et al. (1994), Nature (London), 264:436-440; Nobori, T., et al. (1994), Nature (London), 368:753-756), p27 (see, Polyak, K., et al. (1994), Cell, 78:59-66; Toyoshima, H., et al. (1994), Cell, 78:67-74), p28 (see, Hengst, L., et al. (1994), Proc. Natl. Acad. Sci. U.S.A., 91:5291-5295) and p21 (see, Gu, Y., et al. (1993), Nature (London), 366:707-710; Xiiong, Y., et al. (1993), Nature (London), 366:701-704; Harper, J. W., et al. (1993), Cell, 75:805-816; Dulic, V., et al. (1994), Cell, 76:1013-1023), the latter being inducible by p53. Especially noteworthy is the fact that deletions of the pl6 gene were found in over 50% of all human malignant cell lines tested (see, Kamb, A., supra, Nobori, T., et al., supra), although much less in primary tumor cells (see, Spruck III, C. H., et al. (1994), Nature (London), 370:183-184), implicating p16 functions as tumor suppressor protein. Thus, both the cell growth signals transmitted through many oncogene products and the growth inhibitory signals from several tumor suppressor proteins modulate the activity of CDKs. Although mutations in CDKs themselves have not been associated with cancer, cyclin overexpression has been linked to tumorigenesis (see, Hunter, T., et al. (1991), Cell, 66:1071-1074; Keyomarsi, K., et al. (1993), Proc. Natl. Acad. Sci. U.S.A., 90:1112-1116; Wang, T. C., et al. (1994), Nature (London), 369:669-671.) Hence, CDKs are a promising target for developing inhibitors with antineoplastic effects and for the treatment of cell-proliferative diseases.
The purine ring system is a key structural element of the substrates and ligands of many biosynthetic, regulatory and signal transduction proteins including cellular kinases, G proteins and polymerases. As such, the purine ring system has been a good starting point in the search for inhibitors of many biomedically significant processes. In fact, while purine analogs were being screened for inhibition of various protein kinases, a relatively selective inhibitor, olomoucine (FIG. 1), was identified that competitively inhibits CDK2/cyclin A with an IC.sub.50 of 7 .mu.M (see, Vesely, J., et al., (1994) Eur. J. Biochem., 224:771-786). Although olomoucine exhibits moderate inhibitory activity and good selectivity for the CDK/cyclin protein kinases, it would be advantageous to identify other purine analogs that have increased affinity and specificity for protein kinases as well as G proteins and polymerases. Quite surprisingly, the present invention provides such analogs.